| *** yamamoto has joined #openstack-lbaas | 00:02 | |
| *** slaweq has joined #openstack-lbaas | 00:04 | |
| *** yuanying_ has joined #openstack-lbaas | 00:04 | |
| *** yuanying has quit IRC | 00:05 | |
| rm_work | https://review.openstack.org/#/c/504875/ is really good ... | 00:08 |
|---|---|---|
| rm_work | if we can merge that, it'd be A++ | 00:08 |
| *** slaweq has quit IRC | 00:09 | |
| *** fnaval has quit IRC | 00:14 | |
| *** threestrands has quit IRC | 00:21 | |
| *** fnaval has joined #openstack-lbaas | 00:36 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: WIP: Floating IP Network Driver (spans L3s) https://review.openstack.org/435612 | 00:39 |
| *** sshank has quit IRC | 00:39 | |
| xgerman_ | rm_work: https://review.openstack.org/#/c/523946/2 | 00:44 |
| xgerman_ | will take a break though + check back later | 00:44 |
| rm_work | that works even with the owner being OCTAVIA_OWNER? | 00:44 |
| *** yamamoto has quit IRC | 00:45 | |
| rm_work | I mean, it's fine by me if it works for you | 00:45 |
| rm_work | xgerman_: my driver does it differently anyway so doesn't affect me :P | 00:45 |
| rm_work | but if johnsom tested and it works... | 00:45 |
| rm_work | I am ok to merge | 00:46 |
| johnsom | I am booting a act/stdby just to make sure | 00:46 |
| xgerman_ | k | 00:47 |
| johnsom | yeah, seems fine to me | 00:51 |
| *** fnaval has quit IRC | 00:51 | |
| *** fnaval has joined #openstack-lbaas | 00:51 | |
| *** jniesz has quit IRC | 00:52 | |
| rm_work | k | 00:52 |
| openstackgerrit | Bar RH proposed openstack/octavia master: Fix filtering in list API calls https://review.openstack.org/522689 | 01:06 |
| *** links has joined #openstack-lbaas | 01:09 | |
| *** sanfern has joined #openstack-lbaas | 01:20 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: WIP: Floating IP Network Driver (spans L3s) https://review.openstack.org/435612 | 01:24 |
| *** yamamoto has joined #openstack-lbaas | 01:24 | |
| *** armax has quit IRC | 01:26 | |
| *** yuanying_ has quit IRC | 01:29 | |
| *** yuanying has joined #openstack-lbaas | 01:29 | |
| *** bar_ has quit IRC | 01:31 | |
| *** yuanying_ has joined #openstack-lbaas | 01:33 | |
| *** jcross has quit IRC | 01:34 | |
| *** yuanying has quit IRC | 01:37 | |
| *** annp has joined #openstack-lbaas | 01:56 | |
| sapd_ | rm_work , I still can't understand how amphora-agent send heartbeat to health-manager through UDP, | 02:01 |
| johnsom | sapd_ The code is here: https://github.com/openstack/octavia/tree/master/octavia/amphorae/backends/health_daemon | 02:04 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: WIP: Floating IP Network Driver (spans L3s) https://review.openstack.org/435612 | 02:13 |
| *** yamamoto has quit IRC | 02:22 | |
| *** yamamoto has joined #openstack-lbaas | 02:24 | |
| *** yamamoto has quit IRC | 02:40 | |
| *** sanfern has quit IRC | 02:44 | |
| rm_work | hmmm, need to fix our client to not explode if it doesn't find the network-id you pass in... if it doesn't find it in the list, it needs to just pass it through assuming it's an ID or something | 02:47 |
| rm_work | <_< | 02:47 |
| *** yamamoto has joined #openstack-lbaas | 02:54 | |
| *** threestrands has joined #openstack-lbaas | 02:55 | |
| sapd_ | Thanks because octavia run in docker container, when I run I just expose 5555 TCP port instead of UDP, so amphora-agent can't send heartbeat | 02:58 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: WIP: Floating IP Network Driver (spans L3s) https://review.openstack.org/435612 | 03:04 |
| *** yamamoto has quit IRC | 03:05 | |
| rm_work | ah yeah | 03:06 |
| rm_work | that is important :) | 03:06 |
| rm_work | interestingly, if the health-manager never gets an initial heartbeat for an amphora, it will never try to take failover operations on it | 03:06 |
| rm_work | which could be a bug or a feature <_< | 03:06 |
| rm_work | i've looked at fixing that several times but it's a lot of work for limited gain | 03:06 |
| *** dougwig has quit IRC | 03:18 | |
| openstackgerrit | Merged openstack/octavia master: Updated from global requirements https://review.openstack.org/523768 | 03:20 |
| openstackgerrit | Merged openstack/octavia master: Update devstack plugin and examples https://review.openstack.org/503638 | 03:23 |
| openstackgerrit | Merged openstack/octavia master: Enable some off-by-default checks https://review.openstack.org/522178 | 03:23 |
| *** sticker has quit IRC | 03:58 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 04:15 | |
| *** AlexeyAbashkin has quit IRC | 04:20 | |
| *** tongl has joined #openstack-lbaas | 04:29 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Fix the failover API to not fail with immutable LB https://review.openstack.org/523242 | 04:49 |
| *** armax has joined #openstack-lbaas | 05:31 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: WIP: Floating IP Network Driver (spans L3s) https://review.openstack.org/435612 | 05:34 |
| *** yamamoto has joined #openstack-lbaas | 05:39 | |
| *** yamamoto has quit IRC | 05:43 | |
| *** sanfern has joined #openstack-lbaas | 05:50 | |
| *** yamamoto has joined #openstack-lbaas | 05:53 | |
| *** yamamoto has quit IRC | 05:57 | |
| *** sanfern has quit IRC | 06:00 | |
| *** gcheresh has joined #openstack-lbaas | 06:04 | |
| *** pcaruana has joined #openstack-lbaas | 06:05 | |
| *** armax has quit IRC | 06:07 | |
| *** armax has joined #openstack-lbaas | 06:10 | |
| *** pcaruana has quit IRC | 06:16 | |
| *** eN_Guruprasad_Rn has joined #openstack-lbaas | 06:16 | |
| *** Alex_Staf has joined #openstack-lbaas | 06:20 | |
| *** fnaval has quit IRC | 06:27 | |
| *** yamamoto has joined #openstack-lbaas | 06:31 | |
| *** ipsecguy has quit IRC | 06:40 | |
| *** armax has quit IRC | 06:40 | |
| *** ipsecguy has joined #openstack-lbaas | 06:40 | |
| *** yamamoto has quit IRC | 06:43 | |
| *** armax has joined #openstack-lbaas | 06:48 | |
| openstackgerrit | Merged openstack/octavia master: Fix functional tests for MacOS https://review.openstack.org/523637 | 06:50 |
| *** threestrands has quit IRC | 06:50 | |
| *** yamamoto has joined #openstack-lbaas | 06:59 | |
| *** yamamoto has quit IRC | 07:06 | |
| *** aojea has joined #openstack-lbaas | 07:07 | |
| *** armax has quit IRC | 07:11 | |
| *** Alex_Staf has quit IRC | 07:13 | |
| *** aojea has quit IRC | 07:14 | |
| openstackgerrit | Merged openstack/octavia master: Adds the user's project id to the VIP port creation https://review.openstack.org/523946 | 07:16 |
| *** rcernin has quit IRC | 07:20 | |
| *** tongl has quit IRC | 07:39 | |
| *** yamamoto has joined #openstack-lbaas | 07:49 | |
| sapd_ | johnsom, The octavia-dashboard is not include Pool , so We can't delete pool on horizon. | 07:51 |
| *** yamamoto has quit IRC | 07:51 | |
| *** pcaruana has joined #openstack-lbaas | 07:55 | |
| *** yamamoto has joined #openstack-lbaas | 07:57 | |
| *** slaweq has joined #openstack-lbaas | 08:06 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 08:06 | |
| *** rcernin has joined #openstack-lbaas | 08:07 | |
| *** slaweq has quit IRC | 08:27 | |
| *** slaweq has joined #openstack-lbaas | 08:31 | |
| *** yamamoto has quit IRC | 08:33 | |
| *** yamamoto has joined #openstack-lbaas | 08:40 | |
| *** slaweq has quit IRC | 08:41 | |
| *** AlexeyAbashkin has quit IRC | 08:41 | |
| rm_work | sapd_: errr.. really? I thought it was pretty complete as far as object mapping | 08:42 |
| *** AlexeyAbashkin has joined #openstack-lbaas | 08:42 | |
| rm_work | you can't get to the pools? | 08:42 |
| rm_work | or the delete button just isn't there? | 08:42 |
| *** slaweq has joined #openstack-lbaas | 08:44 | |
| *** yamamoto has quit IRC | 08:44 | |
| sapd_ | It does not show member in pool, It just show default pool ID in listener row. | 08:47 |
| sapd_ | rm_work I think, It has not implemented yet. | 08:48 |
| rm_work | i was pretty sure the whole thing "worked" from end to end, more or less, last i checked | 08:48 |
| rm_work | but definitely CRUD operations for the basic objects necessary for creating a working LB | 08:48 |
| rm_work | can you printscreen what you have and show me on imgur or something? if it's not sensitive data | 08:49 |
| sapd_ | rm_work, here you are: https://imgur.com/a/otXPY | 08:52 |
| sapd_ | I can't find any button about pool or member in pool when listener is created | 08:52 |
| rm_work | errr | 08:53 |
| rm_work | that's octavia-dashboard? | 08:54 |
| rm_work | hmmmm | 08:54 |
| sapd_ | yes. I installed from this source: https://github.com/openstack/octavia-dashboard | 08:54 |
| rm_work | hmm. it looks different than i remember | 08:55 |
| rm_work | i'd need to spin up a stack again | 08:55 |
| nmagnezi | rm_work, o/ | 08:58 |
| rm_work | hey nmagnezi | 08:58 |
| nmagnezi | rm_work, hi Adam | 08:58 |
| nmagnezi | rm_work, not sure why this one is not getting merge https://review.openstack.org/#/c/520863/ | 08:58 |
| nmagnezi | merged* | 08:59 |
| rm_work | child of https://review.openstack.org/#/c/504875/ | 08:59 |
| sapd_ | rm_work I'm waiting | 08:59 |
| rm_work | sapd_: well, that'll take about an hour | 08:59 |
| rm_work | and actually i'm just about to collapse for the night :( | 08:59 |
| *** Alex_Staf has joined #openstack-lbaas | 09:00 | |
| rm_work | nmagnezi: do you have a stack up that you could throw the octavia-dashboard on? | 09:00 |
| rm_work | the screenshots sapd_ linked seem odd to me | 09:00 |
| nmagnezi | rm_work, maybe, lemme check | 09:00 |
| rm_work | but i don't actually use the dashboard much... | 09:00 |
| nmagnezi | rm_work, same here. checking. | 09:01 |
| rm_work | I just remember reviewing it once when michael did the huge conversion and making sure it worked end to end | 09:01 |
| dayou | sapd_, On the third image you posted | 09:01 |
| dayou | You can click the default pool id link | 09:01 |
| dayou | To get to the default pool page | 09:01 |
| rm_work | ah, yes | 09:01 |
| dayou | Anyway, current dashboard is based on quite old code that was developed about 2 years ago | 09:02 |
| dayou | I am working on rebuild it based on the latest horizon framework, still took a few days to get the initial work done | 09:02 |
| *** sri_ has joined #openstack-lbaas | 09:03 | |
| dayou | But would be in this week | 09:03 |
| rm_work | ah, nice | 09:03 |
| nmagnezi | rm_work, k. I have it running. what do you wish me to check? if it's possible to delete a pool? | 09:03 |
| rm_work | i think maybe dayou solved it, but | 09:04 |
| rm_work | does the main loadbalancer page really look like that? https://imgur.com/a/otXPY | 09:04 |
| *** yamamoto has joined #openstack-lbaas | 09:04 | |
| nmagnezi | rm_work, yes | 09:12 |
| sapd_ | oh yes! | 09:13 |
| sapd_ | I got it. dayou | 09:14 |
| nmagnezi | nevertheless, I do see pools | 09:14 |
| nmagnezi | and I do see where do delete them | 09:14 |
| nmagnezi | s/do/to | 09:14 |
| rm_work | hmmmm | 09:14 |
| rm_work | well, good :P | 09:15 |
| nmagnezi | :D | 09:15 |
| rm_work | probably needs some work from a usability standpoint I guess T_T | 09:15 |
| rm_work | which is what dayou is doing? :P | 09:15 |
| nmagnezi | yeah I totally agree it can and should be improved | 09:16 |
| nmagnezi | but for what it's worth, it's there. | 09:16 |
| sapd_ | I think when click to Listener, We should show Default pool member page. | 09:16 |
| *** yamamoto has quit IRC | 09:19 | |
| *** eN_Guruprasad_Rn has quit IRC | 09:19 | |
| *** eN_Guruprasad_Rn has joined #openstack-lbaas | 09:19 | |
| dayou | https://storyboard.openstack.org/#!/story/1713855 | 09:21 |
| dayou | Michael actually reported a bug on this | 09:22 |
| dayou | https://pasteboard.co/GW1fC5Ho.png | 09:22 |
| dayou | Here is what it looks like in the version I am developing, just for a preview | 09:22 |
| dayou | To adress this bug | 09:23 |
| bcafarel | nmagnezi: working on merging amphora-agent elements, I noticed in rdo packaging we do not have agent log rotation ( somehting similar to https://github.com/openstack/octavia/blob/master/elements/amphora-agent/static/etc/logrotate.d/amphora-agent ), will send a review there to fix | 09:31 |
| *** bar_ has joined #openstack-lbaas | 09:34 | |
| sapd_ | dayou, good job :D | 09:37 |
| dayou | Hopefully it could be improved by the queens cycle, still a lot of work to do ahead | 09:39 |
| nmagnezi | bcafarel, ack. :) | 09:39 |
| sapd_ | https://storyboard.openstack.org/#!/story/2001374 | 09:39 |
| *** pcaruana has quit IRC | 09:39 | |
| dayou | sapd_, thanks for the encouragement, anyway! | 09:39 |
| sapd_ | Has it fixed? | 09:39 |
| *** pcaruana has joined #openstack-lbaas | 09:40 | |
| dayou | Not yet, has to fix the bug one by one | 09:40 |
| dayou | Currently I just try to get the base done | 09:41 |
| *** fnaval has joined #openstack-lbaas | 09:43 | |
| *** kbyrne has quit IRC | 09:43 | |
| *** kbyrne has joined #openstack-lbaas | 09:44 | |
| dayou | Already on a 16x7 mode, too much work to be done. | 09:44 |
| sapd_ | :D | 09:45 |
| sapd_ | I think so | 09:46 |
| *** fnaval has quit IRC | 09:48 | |
| cgoncalves | bar_: o/ | 09:58 |
| bar_ | cgoncalves, hey! | 09:58 |
| cgoncalves | bar_: re your comment on https://review.openstack.org/#/c/518767/8/octaviaclient/api/load_balancer_v2.py | 09:58 |
| bar_ | cgoncalves, what about it? | 09:58 |
| cgoncalves | bar_: I'm thinking of submitting a patch dedicated to that because I also need to add :raises for exceptions | 09:58 |
| cgoncalves | :return: A ``dict`` representing the updated resource | 09:59 |
| cgoncalves | :raises: $some_msg_mentioning the possible outcomes (HTTP 400-500) | 10:00 |
| bar_ | What does it raises? | 10:00 |
| cgoncalves | bar_: https://developer.openstack.org/api-ref/load-balancer/v2/ | 10:00 |
| cgoncalves | bar_: check the 'Error' table for earch API call | 10:01 |
| bar_ | is an error response (e.g. code 400) raises an exception from these functionsL | 10:02 |
| bar_ | ? | 10:02 |
| bar_ | specifically those who don't have the correct_return_codes decorator. | 10:03 |
| bar_ | (I'm asking you, I do not remember) | 10:04 |
| cgoncalves | bar_: for the ones with decorators, yes. I need to check for the others | 10:06 |
| bar_ | IMO, since all other functions in that file lack the :raises: comment, you shouldn't introduce it in your quota patch, unless one of your functions has some unique behavior you are aware of | 10:07 |
| cgoncalves | bar_: agreed, hence my suggestion to address your comment in a separate patch + adding :raises | 10:09 |
| *** xgerman_ has quit IRC | 10:09 | |
| *** fyxim has quit IRC | 10:09 | |
| *** kong has quit IRC | 10:09 | |
| *** LutzB has quit IRC | 10:09 | |
| bar_ | cgoncalves, why separate patch for the :return:? (it was not my only comment) | 10:10 |
| cgoncalves | bar_: I would leave all :return's as-is including the one you commented on to be consistent with the rest of the :return's | 10:11 |
| cgoncalves | bar_: then on a new patch I'd fix the :return for all API and add :raises | 10:11 |
| cgoncalves | because current :returns say "Response Code from API" which is not true. the return is a dict of the resource (when there's a body) | 10:12 |
| bar_ | cgoncalves, regarding the :return:, you shouldn't introduce mistakes in a patch, and it should be corrected within the same patch. This case differs from adding :raises:, because there the motivation is *lack* of documentary, not *false* documentary | 10:14 |
| *** xgerman_ has joined #openstack-lbaas | 10:15 | |
| *** fyxim has joined #openstack-lbaas | 10:15 | |
| *** kong has joined #openstack-lbaas | 10:15 | |
| *** LutzB has joined #openstack-lbaas | 10:15 | |
| cgoncalves | bar_: ok, then I'd argue the same for the loadbalancer vs load-balancer comment you made ;-) | 10:17 |
| cgoncalves | https://review.openstack.org/#/c/518767/8/octaviaclient/osc/v2/quota.py | 10:17 |
| bar_ | cgoncalves, anyway, it's a great patch, you should perfect it so it can be included in the coming Q2 milestone. | 10:18 |
| bar_ | cgoncalves, you could argue it's nitty, but not the same, since 'raises' is documentation extension, and CLI command is actually something we use. | 10:20 |
| *** annp has quit IRC | 10:23 | |
| cgoncalves | ok. thanks | 10:26 |
| bar_ | np. perhaps you could take the time to review my patch? https://review.openstack.org/#/c/522666/ | 10:27 |
| cgoncalves | bar_: certainly, will do! | 10:33 |
| cgoncalves | bar_: just by reading the commit message, I tend to not agree with commits adding bunch of different stuff | 10:34 |
| bar_ | cgoncalves, yeah, I will most probably have to break it in the end, there're too many dependencies anyway | 10:35 |
| cgoncalves | bar_: +1 | 10:35 |
| cgoncalves | to that comment, not to the patch :P | 10:35 |
| * cgoncalves mumbles 'why, oh why quota list and quota show' :S | 10:40 | |
| cgoncalves | nmagnezi: what's your stance on https://review.openstack.org/#/c/518767/8/octaviaclient/osc/v2/quota.py@121 ? | 10:45 |
| cgoncalves | nmagnezi: in case I reverse to '--loadbalancer' you would need to update your patch | 10:46 |
| nmagnezi | cgoncalves, I was not following, in the middle of a review. will look a bit later | 10:46 |
| nmagnezi | cgoncalves, generally i would say that all attribute names that consist of more than one word, those words should be separated with a hyphen. that's on par (and consistent) with the general guidelines. that being said i'm not sure how to look at the 'loadbalancer' word that comes right after the 'openstack' word in CLI. I *think it's kinda different. | 10:51 |
| *** sri_ has quit IRC | 11:09 | |
| openstackgerrit | Hengqing Hu proposed openstack/octavia-dashboard master: Provide default translation for en_GB locale https://review.openstack.org/524139 | 11:09 |
| *** Dave has quit IRC | 11:12 | |
| *** Dave has joined #openstack-lbaas | 11:15 | |
| cgoncalves | nmagnezi: k | 11:16 |
| openstackgerrit | Carlos Goncalves proposed openstack/python-octaviaclient master: Add Quota client API and OSC support https://review.openstack.org/518767 | 11:19 |
| bar_ | cgoncalves, Are you aware to the fact the Nir's patch is in-line with your change ('loadbalancer' --> 'load-balancer')? | 11:28 |
| cgoncalves | bar_: yes: https://review.openstack.org/#/c/521482/ | 11:30 |
| cgoncalves | bar_: thus my comment that either I leave it as is ('--load-balancer') or Nir needs to update his patch later once mine gets merged | 11:31 |
| cgoncalves | otherwise we end up again with a mix of loadbalancer (for quota) and load-balancer (the rest) :) | 11:31 |
| bar_ | cgoncalves, your patch will unlikely merge after his. | 11:32 |
| cgoncalves | bar_: exactly. so I'm not getting your point, sorry :) | 11:33 |
| bar_ | cgoncalves, You cannot expect his patch to revert yours, because his patch is in-line with yours. | 11:34 |
| openstackgerrit | Hengqing Hu proposed openstack/octavia-dashboard master: Catch up with horizon framework https://review.openstack.org/523249 | 11:34 |
| bar_ | so should he abandon the change ('loadbalancer' --> 'load-balancer'), which in my view - he should. It is not his patch's place to revert yours. | 11:35 |
| cgoncalves | bar_: k, you're right in the sense his patch does not revert mine | 11:36 |
| cgoncalves | yet loadbalancer -> load-balancer renaming is needed and he's addressing for the rest | 11:37 |
| bar_ | cgoncalves, this is very well, but you cannot argue that Nir's patch will revert yours - it won't, whether his proposal is accepted or not. | 11:39 |
| cgoncalves | bar_: yes, you're right | 11:40 |
| cgoncalves | bar_: what I meant to say was that your comment on the review was to have --loadbalancer. I slightly tend to disagree because we would be commiting something that we know beforehand does not follow the CLI guidelines | 11:41 |
| cgoncalves | still, if I change to --loadbalancer, Nir would have to update his patch to s/loadbalancer/load-balancer on the quota.py file | 11:42 |
| bar_ | and it would make sense, because Nir's patch is exactly that - meant to standardize the whole client, not feature specific like yours. | 11:43 |
| *** fnaval has joined #openstack-lbaas | 11:43 | |
| cgoncalves | bar_: ok, I'm okay doing so. I'll follow closely Nir's patch to make sure --load-balancer in quota is not forgotten ;-) | 11:44 |
| bar_ | cgoncalves, very well. I guess I'll have to argue against it in both fronts. | 11:45 |
| bar_ | lol | 11:45 |
| cgoncalves | bar_: can we agree on something? :D | 11:45 |
| bar_ | cgoncalves, if youre in the right spirit, you know you could support both options easy, right? | 11:46 |
| *** fnaval has quit IRC | 11:47 | |
| cgoncalves | bar_: yes but I wouldn't do that anyway :P | 11:50 |
| cgoncalves | bar_: ah, plus with the renaming we also need to make sure it's backward compatible for a period of time | 11:51 |
| *** rcernin has quit IRC | 11:52 | |
| cgoncalves | the odds of that we can agree that the Earth is round are very high -- don't tell me otherwise, pleeeease ;D | 11:52 |
| *** bar_ has quit IRC | 11:59 | |
| *** bar_ has joined #openstack-lbaas | 12:01 | |
| *** salmankhan has joined #openstack-lbaas | 12:15 | |
| *** yamamoto has joined #openstack-lbaas | 12:37 | |
| *** bar_ has quit IRC | 12:51 | |
| *** bar_ has joined #openstack-lbaas | 12:57 | |
| *** atoth has joined #openstack-lbaas | 13:00 | |
| *** sanfern has joined #openstack-lbaas | 13:01 | |
| sanfern | Hi johnsom, | 13:02 |
| *** yamamoto has quit IRC | 13:04 | |
| *** yamamoto has joined #openstack-lbaas | 13:05 | |
| *** sanfern has quit IRC | 13:37 | |
| *** fnaval has joined #openstack-lbaas | 13:43 | |
| *** fnaval has quit IRC | 13:47 | |
| *** links has quit IRC | 13:54 | |
| *** sanfern has joined #openstack-lbaas | 13:56 | |
| nmagnezi | bar_, openstack port create mytestport2 --network private --security-group 8e71e599-261f-4861-a661-492d03b9b13b | 13:56 |
| *** leitan has joined #openstack-lbaas | 14:04 | |
| *** leitan has quit IRC | 14:09 | |
| *** salmankhan has quit IRC | 14:18 | |
| openstackgerrit | Carlos Goncalves proposed openstack/python-octaviaclient master: Add Quota client API and OSC support https://review.openstack.org/518767 | 14:20 |
| openstackgerrit | Bar RH proposed openstack/octavia master: Fail-proof VIP deallocation task https://review.openstack.org/523931 | 14:33 |
| *** Swami has joined #openstack-lbaas | 14:42 | |
| *** armax has joined #openstack-lbaas | 14:50 | |
| *** fnaval has joined #openstack-lbaas | 14:54 | |
| *** fnaval has quit IRC | 14:58 | |
| *** sanfern has quit IRC | 15:04 | |
| *** eN_Guruprasad_Rn has quit IRC | 15:07 | |
| *** fnaval has joined #openstack-lbaas | 15:17 | |
| -openstackstatus- NOTICE: if you receieved a result of "RETRY_LIMIT" after 14:15 UTC, it was likely due to an error since corrected. please "recheck" | 15:37 | |
| *** isp is now known as isanto | 15:47 | |
| *** isanto is now known as isantosp | 15:47 | |
| *** isantosp is now known as isp | 15:47 | |
| *** isp is now known as issp | 15:47 | |
| *** yamamoto has quit IRC | 15:49 | |
| *** yamamoto has joined #openstack-lbaas | 15:50 | |
| *** slaweq has quit IRC | 15:52 | |
| *** eN_Guruprasad_Rn has joined #openstack-lbaas | 15:52 | |
| cgoncalves | xgerman_, rm_work, johnsom: if you have some review cycles: https://review.openstack.org/#/c/521250/ | 16:02 |
| *** eN_Guruprasad_Rn has quit IRC | 16:08 | |
| *** Alex_Staf has quit IRC | 16:51 | |
| *** sanfern has joined #openstack-lbaas | 16:57 | |
| *** AlexeyAbashkin has quit IRC | 17:01 | |
| sanfern | Hi johnsom | 17:07 |
| johnsom | sanfern o/ | 17:07 |
| *** pcaruana has quit IRC | 17:08 | |
| sanfern | Amphora service_type - SERVICE_TYPE_DISTRIBUTOR, SERVICE_TYPE_DISTRIBUTOR_LOADBALANCER and SERVICE_TYPE_LOADBALANCER - what do you say ? | 17:10 |
| johnsom | Is "SERVICE_TYPE_DISTRIBUTOR_LOADBALANCER" meaning it runs HAProxy and the distributor on the same instance? | 17:10 |
| sanfern | yes sir | 17:11 |
| *** jcross has joined #openstack-lbaas | 17:11 | |
| johnsom | Not sure we are going there any time soon, I would probably start with the other two. | 17:11 |
| sanfern | ok | 17:11 |
| *** Swami has quit IRC | 17:12 | |
| sanfern | As per l3-active-active spec in amphora_registration table - load_balancer_id is made nullable=False then it should be changed to True | 17:14 |
| sanfern | Spec needs fix | 17:14 |
| johnsom | Ummm, we usually do that to allow a smooth upgrade path. Let me go re-read it | 17:15 |
| johnsom | Ah, yeah, I see. You are correct | 17:16 |
| johnsom | Well, this is mapping table right? What is the case we would add a record here where there is no load balancer ID? | 17:17 |
| sanfern | it will not allow insert | 17:20 |
| sanfern | it is a mapping table | 17:21 |
| sanfern | johnsom, Can we generate alembic_migration script file without devstack ? | 17:23 |
| johnsom | Yes, it's just a python tool | 17:25 |
| *** salmankhan has joined #openstack-lbaas | 17:31 | |
| sanfern | johnsom, how please ? | 17:35 |
| sanfern | how do I get Revision ID: ? | 17:36 |
| johnsom | http://alembic.zzzcomputing.com/en/latest/tutorial.html#create-a-migration-script | 17:36 |
| sanfern | my mac do not have this tool :( | 17:39 |
| johnsom | You should be able to pip install it | 17:40 |
| sanfern | oops thanks johnsom | 17:42 |
| *** SumitNaiksatam has joined #openstack-lbaas | 17:53 | |
| sanfern | johnsom, help please - http://paste.openstack.org/show/bXvHo8AJdhDpxmkAhtqU/ | 17:55 |
| johnsom | Oh, right, we changed that out. | 17:59 |
| johnsom | octavia-db-manage --config-file octavia/db/migration/alembic.ini revision -m "create distributor table" | 17:59 |
| johnsom | There is a wrapper script for alembic | 18:00 |
| johnsom | sanfern ^^^ | 18:00 |
| sanfern | ok thanks johnsom | 18:00 |
| *** openstackgerrit has quit IRC | 18:48 | |
| *** pcaruana has joined #openstack-lbaas | 18:50 | |
| *** slaweq has joined #openstack-lbaas | 18:51 | |
| *** slaweq has quit IRC | 18:53 | |
| *** openstackgerrit has joined #openstack-lbaas | 18:55 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Fix the failover API to not fail with immutable LB https://review.openstack.org/523242 | 18:55 |
| bar_ | johnsom, +1 | 18:56 |
| johnsom | Grin, yeah, too many things going on at once. At least I had the reference tag in there | 18:56 |
| bar_ | hehe | 18:57 |
| *** bar_ has quit IRC | 19:25 | |
| *** bar_ has joined #openstack-lbaas | 19:29 | |
| openstackgerrit | Bar RH proposed openstack/python-octaviaclient master: Complement Octavia client with a set of features https://review.openstack.org/522666 | 19:50 |
| *** AlexeyAbashkin has joined #openstack-lbaas | 19:52 | |
| *** AlexeyAbashkin has quit IRC | 19:59 | |
| *** SumitNaiksatam has quit IRC | 20:00 | |
| *** bar_ has quit IRC | 20:06 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 20:12 | |
| *** AlexeyAbashkin has quit IRC | 20:17 | |
| *** slaweq has joined #openstack-lbaas | 20:32 | |
| *** slaweq has quit IRC | 20:34 | |
| *** slaweq has joined #openstack-lbaas | 20:34 | |
| openstackgerrit | Jude Cross proposed openstack/neutron-lbaas master: [WIP] Remove unnecessary lazy-loaded queries https://review.openstack.org/477698 | 20:50 |
| *** pcaruana has quit IRC | 20:55 | |
| *** gcheresh has quit IRC | 20:58 | |
| *** gcheresh has joined #openstack-lbaas | 21:03 | |
| *** threestrands has joined #openstack-lbaas | 21:12 | |
| *** threestrands has joined #openstack-lbaas | 21:12 | |
| *** pcaruana has joined #openstack-lbaas | 21:24 | |
| *** sshank has joined #openstack-lbaas | 21:29 | |
| *** salmankhan has quit IRC | 21:38 | |
| *** pcaruana has quit IRC | 21:39 | |
| *** bar_ has joined #openstack-lbaas | 21:41 | |
| *** gcheresh has quit IRC | 21:57 | |
| *** fnaval has quit IRC | 22:13 | |
| *** rcernin has joined #openstack-lbaas | 22:18 | |
| *** rcernin has quit IRC | 22:21 | |
| *** rcernin has joined #openstack-lbaas | 22:21 | |
| *** sanfern has quit IRC | 22:38 | |
| *** sanfern has joined #openstack-lbaas | 22:39 | |
| jcross | Hey @johnsom I am looking at some lazy load sqlalchemy stuff for patch set: https://review.openstack.org/#/c/477698/14 basically tests dont pass because the DB replies without having the listener actually populated after lazy is disabled. By any chance do you know where to start with this? | 22:42 |
| jcross | this is for neutron-lbaas btw :P | 22:43 |
| jcross | my sqlalchemy foo is terrible. | 22:43 |
| jcross | essentially this is what I am seeing: https://pastebin.com/anrQNNpp | 22:45 |
| johnsom | hmmm | 22:49 |
| openstackgerrit | Merged openstack/octavia master: Split amphora haproxy systemd unit https://review.openstack.org/520629 | 22:49 |
| jcross | seems to me like a race condition caused by the lazy loading but I am not sure where to start to fix that :/ | 22:52 |
| jcross | rather... caused by the lazy loading no longer being there. | 22:52 |
| johnsom | Well, my *guess* (I am totally context switching here...) is that there is no defined relationship between the pool and the listeners like there is for healthmonitor, session_persistence, and loadbalancer. So without the "joined" lazy load there is no longer an implied relationship because there is a relationship in the DB. | 22:56 |
| johnsom | I don't have a nlbaas VM booted at the moment to confirm | 22:56 |
| jcross | no problem I am just looking for direction thank you :) | 22:57 |
| bar_ | johnsom, hey | 22:59 |
| johnsom | bar_ Hi | 22:59 |
| bar_ | I'm looking into Quota API, I want to lose the project_id requirement. and it's pretty straight forward to accept 'null' instead of valid project_id and infer it from the session context | 23:01 |
| *** slaweq has quit IRC | 23:02 | |
| bar_ | is it cool to leave the API as it is and accept '/v2.0/lbaas/quotas/null', or change/add-to the api to accept /v2.0/lbaas/quotas ? | 23:02 |
| johnsom | jcross Yeah, that model just seems sketchy for pool.listeners. There is even a comment that it is sketchy in the main code therre "# sqlalchemy cries if listeners is defined.". Basically I think they have been living on an implied model extension for the listeners, but the actual model doesn't have it. | 23:02 |
| *** slaweq has joined #openstack-lbaas | 23:02 | |
| johnsom | bar_ hmmm. This is for the show, update, and reset actions? | 23:05 |
| openstackgerrit | Carlos Goncalves proposed openstack/python-octaviaclient master: Add Quota client API and OSC support https://review.openstack.org/518767 | 23:06 |
| bar_ | basically, yeah | 23:06 |
| bar_ | the broader question is, how committed are we to the API in its current state? | 23:07 |
| *** slaweq has quit IRC | 23:07 | |
| johnsom | I vaguely remember this conversation, but what is the use case for making these calls with the requester project_id? There is no real usecase where a tenant user would change their own quota right? | 23:07 |
| johnsom | Right now it's an Admin only API | 23:07 |
| bar_ | aw, for how long? | 23:08 |
| bar_ | Because there is a use-case for show, for non-admin tenant. | 23:08 |
| johnsom | Yeah, show I get. PUT and DELETE not. | 23:09 |
| bar_ | right | 23:09 |
| johnsom | Ok, getting a context switch through here.... | 23:09 |
| bar_ | perhaps another time then? | 23:09 |
| johnsom | So you want a show quota that doesn't have a project_id in it? That is List.... | 23:10 |
| johnsom | I mean overall the API convention in OpenStack is "show" has an ID in the path right? | 23:10 |
| johnsom | Now is ok. I just need to switch. I was in the provider driver spec, pivoted to the data models in nlbaas, now back to octavia quotas.... Just takes me a minute | 23:11 |
| bar_ | :-) | 23:11 |
| bar_ | I believe Quota is a bit different, because it is popular use case to call for quota_show() on yourself, don't you agree? | 23:12 |
| *** AlexeyAbashkin has joined #openstack-lbaas | 23:12 | |
| bar_ | and that is different than quota_list(), because you want to get the state of self. | 23:12 |
| bar_ | only. | 23:13 |
| johnsom | Your end goal is in OSC to be able to have a show quota command that doesn't require a parameter right? | 23:13 |
| bar_ | exactly | 23:13 |
| johnsom | Well, list with permissions to only see your project implies a list of one.... | 23:14 |
| johnsom | Ok, I think I am caught up to you | 23:15 |
| bar_ | either have it all (admin) or fend for yourself? | 23:15 |
| johnsom | This is probably why some commands have --all parameters for list | 23:15 |
| bar_ | there's also a case with show member, it requires pool_id. isn't member_id 1:1? | 23:16 |
| *** AlexeyAbashkin has quit IRC | 23:16 | |
| johnsom | name isn't | 23:16 |
| bar_ | GET /v2.0/lbaas/pools/{pool_id}/members/{member-id} | 23:17 |
| johnsom | There is also the concept of shared members I think | 23:17 |
| johnsom | So, the ID could be the same. I don't think we support that in octavia today | 23:18 |
| johnsom | We would have to go back to the lbaasv2 API spec. It pre-dates me a bit | 23:18 |
| bar_ | lbaasv2 == octaviav2? | 23:18 |
| johnsom | octaviav2 is a superset of lbaasv2 for API compatibility | 23:19 |
| bar_ | aha | 23:19 |
| johnsom | Quotas however is one of those extensions Octavia made | 23:19 |
| bar_ | would you agree with GET /v2.0/lbaas/pools/{pool_id}/members/{member-id} --> GET /v2.0/members/{member-id} , had I proposed it? | 23:20 |
| johnsom | So, looking at our neighbor project, neutron: https://developer.openstack.org/api-ref/network/v2/index.html#quotas-extension-quotas | 23:20 |
| johnsom | bar_ It's something I would have to think about. The odd thing there is you can't have a member without a pool, so it would only be gets | 23:21 |
| johnsom | It's an odd data model thing | 23:21 |
| bar_ | also: PUT /v2.0/lbaas/pools/{pool_id}/members/{member_id} -> PUT /v2.0/octavia/members/{member_id} | 23:22 |
| bar_ | and DELETE too.... | 23:23 |
| johnsom | bar_ Right now you cannot do that. The data model doesn't have a concept of just a member. | 23:23 |
| bar_ | I understand, I didn't mean create, but update call | 23:24 |
| johnsom | So, let's focus on one discussion at a time... Why don't we bring up the other API changes at a weekly meeting. | 23:25 |
| bar_ | sure | 23:26 |
| bar_ | thx | 23:26 |
| johnsom | I will say, we cannot drop the old way of accessing those. It would be a major revision roll of the api... | 23:26 |
| bar_ | should we present new api for members, will it be lbaas or Octavia? | 23:27 |
| johnsom | Octavia, there are no new features going into nlbaas | 23:27 |
| bar_ | great, thank you. | 23:28 |
| bar_ | between, I think the client Quota patch is ready for prime time | 23:28 |
| johnsom | So, right now, the way the quota API is setup, you would have a conflict with list and show if you wanted just a show with no project-id. | 23:28 |
| johnsom | Looking at neutron, they don't even have a true "show" | 23:29 |
| johnsom | designate is basically the same as ours: https://developer.openstack.org/api-ref/dns/#quotas | 23:30 |
| johnsom | Ah, no, they don't have a multi-project list option at all | 23:30 |
| johnsom | Ah, they are doing a header trick for list all. hmmm | 23:31 |
| bar_ | yeah, it's quite different from lbaas take of qos | 23:31 |
| johnsom | So, my question to you is are we over thinking this? The client can just fill this in and go on with life. Same with the SDK. Everyone doing a show will have the project ID to fill in since they have to go to keystone. If it's noauth=True we can't imply the project anyway. | 23:32 |
| bar_ | johnsom, yeah, I guess you're right, when it comes to Quota | 23:33 |
| johnsom | This was an issue I ran into with reviewing this quota stuff. There is no standard OpenStack way... | 23:34 |
| bar_ | there is something under openstackclient | 23:34 |
| bar_ | https://github.com/openstack/python-openstackclient/blob/master/openstackclient/common/quota.py | 23:35 |
| bar_ | When it comes to members API, do you see the benefit of accessing a member without the need for pool_id? | 23:37 |
| johnsom | Some characters? Grin | 23:37 |
| bar_ | haha oh well | 23:38 |
| bar_ | at least I cannot be blamed for not trying... | 23:38 |
| johnsom | I am digging for the original spec to see if there is a reason | 23:38 |
| johnsom | Well, don't give up if you see value. It's the community's vote, not just mine.... | 23:39 |
| johnsom | It is here BTW: http://specs.openstack.org/openstack/neutron-specs/specs/kilo/lbaas-api-and-objmodel-improvement.html | 23:41 |
| bar_ | thx | 23:41 |
| johnsom | Some of it probably has to do with the "typical" load balancer model that makes these hierarchical | 23:43 |
| bar_ | Do you think shared-members is valuable feature for octavia? | 23:44 |
| johnsom | Ok, so on quotas we are good? For the members thing, my recommendation is to bring it up that weekly meeting (put it on the agenda) and we can discuss it. | 23:44 |
| bar_ | can you share the like? | 23:45 |
| bar_ | *link | 23:45 |
| johnsom | I think there probably is an optimal reorganization. Fundamentally we don't want multiple health monitor hits per member, so sharing has some value. However the current code/model is not ready for that. | 23:47 |
| johnsom | https://wiki.openstack.org/wiki/Octavia/Weekly_Meeting_Agenda#Meeting_2017-12-06 | 23:47 |
| johnsom | We add agenda items by adding a row before "Open Discussion" with your topic and in parenthesis your IRC handle. | 23:48 |
| johnsom | These are also thing that can go on the roadmap. With our limited resources we do need to focus on our goals for the release (typically decided at the PTGs now), so as a community we do need to balance what we decided on for the release and new stuff. | 23:50 |
| *** sshank has quit IRC | 23:52 | |
| bar_ | johnsom, just added the members topic to the agenda. | 23:53 |
| johnsom | +1 | 23:53 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
